Method and system for overcoming tagging shortage for machine learning purposes

Abstract

A method and a system for overcoming tagged automotive data shortage using machine learning. The system may be implemented over an online automotive data marketplace and may include: a data processing module implemented on a computer processor which receives a set of tagging functions and required automotive data features; a dataset generator implemented on said computer processor which creates custom training sets for a plurality of machine learning algorithms, without tagged data, based on processed automotive data; and a machine learning algorithm generator implemented on said computer processor which trains the machine learning algorithms using the custom training sets, to yield a trained model.

1 . A method of overcoming tagged automotive data shortage, using machine learning, the method comprising:
 receiving a set of tagging functions and required automotive data features;   using machine learning algorithms capable of operating without tagged data based on processed automotive data, to create custom training sets; and   training a supervised model using the custom training sets, to yield a trained model.   
     
     
         2 . The method according to  claim 1 , wherein the set of tagging functions and the required data features are provided by a data consumer. 
     
     
         3 . The method according to  claim 2 , further comprising validating the machine learning algorithms using a tagged dataset from the data consumer. 
     
     
         4 . The method according to  claim 3 , further comprising receiving feedback from the data consumer responsive to the validating of the machine learning algorithms, and repeating the training of the machine learning algorithms, considering said feedback, to yield an improved trained model. 
     
     
         5 . The method according to  claim 1 , wherein the processed automotive data complies with privacy regulations. 
     
     
         6 . The method according to  claim 1 , wherein the tagging functions contextually associate automotive data features with automotive data use cases. 
     
     
         7 . A system for overcoming tagged automotive data shortage using machine learning, the system comprising:
 a data processing module implemented on a computer processor which collects automotive data from data providers and yields processed automotive data;   a dataset generator implemented on said computer processor which received receives a set of tagging functions and required automotive data features and creates custom training sets for a plurality of machine learning algorithms, without tagged data, based on the processed automotive data; and   a machine learning algorithm generator implemented on said computer processor which trains the machine learning algorithms using the custom training sets, to yield a trained model.   
     
     
         8 . The system according to  claim 1 , wherein the set of tagging functions and the required data features are provided by a data consumer. 
     
     
         9 . The system according to  claim 8 , further comprising a consumer hidden evaluation set configured to validate the machine learning algorithms over said computer processor. 
     
     
         10 . The system according to  claim 9 , further wherein said machine learning algorithm generator receives feedback from the data consumer responsive to the validating of the machine learning algorithms, and repeats the training of the machine learning algorithms, considering said feedback, to yield an improved trained model. 
     
     
         11 . The system according to  claim 7 , wherein the processed automotive data complies with privacy regulations. 
     
     
         12 . The system according to  claim 7 , wherein the tagging functions contextually associate automotive data features with automotive data use cases. 
     
     
         13 . A non-transitory computer readable storage medium for overcoming tagged automotive data shortage using machine learning, the computer readable storage medium comprising a set of instructions that when executed cause at least one computer processor to:
 receive a set of tagging functions and required automotive data features;   create custom training sets for a plurality of machine learning algorithms, without tagged data, based on processed automotive data; and   train the machine learning algorithms using the custom training sets, to yield a trained model.   
     
     
         14 . The non-transitory computer readable storage medium according to  claim 13 , wherein the set of tagging functions and the required data features are provided by a data consumer. 
     
     
         15 . The non-transitory computer readable storage medium according to  claim 14 , further comprising computer readable storage medium comprising a set of instructions that when executed cause the at least one computer processor to validate the machine learning algorithms using a tagged dataset from the data consumer. 
     
     
         16 . The non-transitory computer readable storage medium according to  claim 15 , further comprising computer readable storage medium comprising a set of instructions that when executed cause the at least one computer processor to receive feedback from the data consumer responsive to the validating of the machine learning algorithms, and repeating the training of the machine learning algorithms, considering said feedback, to yield an improved trained model. 
     
     
         17 . The non-transitory computer readable storage medium according to  claim 13 , wherein the processed automotive data complies with privacy regulations. 
     
     
         18 . The non-transitory computer readable storage medium according to  claim 13 , wherein the tagging functions contextually associate automotive data features with automotive data use cases.